Spinal Cord (1999) 37, 183 ± 190
ã
1999 International Medical Society of Paraplegia All rights reserved 1362 ± 4393/99 $12.00
http://www.stockton-press.co.uk/sc
Correlates of stress in long-term spinal cord injury
KA Gerhart*,1, DA Weitzenkamp1, P Kennedy2, CA Glass3 and SW Charlifue1
Craig Hospital, Englewood, Colorado, USA; 2National Spinal Injuries Centre, Stoke Mandeville Hospital,
Aylesbury, England; 3Northwest Regional Spinal Injuries Unit, Southport, England
1
Study Design: Longitudinal.
Objectives: To characterize long-injured SCI persons with high reported stress; to assess the
relationship between severity of disability and perceived stress; to identify correlates of future
stress and outcomes of previous stress.
Setting: Two SCI centres in England: Stoke Mandeville Hospital in Aylesbury, and the
District General Hospital in Southport.
Methods: In 1990, 1993 and 1996 187 persons who sustained spinal cord injuries prior to
1971 underwent comprehensive physical evaluations and health status interviews and
completed a battery of tests to measure psychosocial functioning. Using mean scores on the
Perceived Stress Scale (PSS) as the reference, a range of outcomes were analyzed to identify
concurrent, previous, and future variables that were signi®cantly correlated with perceived
stress.
Results: No associations were found between stress and any of the proxy variables that
represented injury severity. Such common SCI-related medical conditions as pressure sores
and upper extremity pain were not related to stress; not even fatigue was signi®cantly
associated with stress in both time periods studied. However, depressive symptoms, poorer life
satisfaction, and poorer perceived well being were associated with future stress and were
outcomes that appeared to be related to earlier stress.
Conclusion: Perceived stress in long-term SCI is not closely related to the severity of the
disability or physical independence. It is, however, related to scores on several measures of
adjustment and coping. Though mean stress scores in this sample did not appear to di€er
substantially from scores in the general nondisabled population, further controlled study is
needed to de®nitively answer the question: Do SCI survivors report more stress than their
nondisabled counterparts?
Sponsorship: This research was supported by a Rehabilitation Research and Training Center
on Aging with Spinal Cord Injury (1993 ± 1997), which was funded by the National Institute
on Disability and Rehabilitation Research of the US Department of Education under grant
number H133B30040. The opinions contained in this publication are those of the authors and
no not necessarily represent those of the US Department of Education.
Keywords: spinal cord injury; stress
Introduction
Stress is both a stimulus and a response, a cause and a
result.1 Indeed, it can be a stimulus for such adverse
outcomes as heart disease, cancer, hypertension,
migraine headaches, peptic ulcers, allergies and
asthma, rheumatoid arthritis, skin conditions, and
such psychological issues as adjustment problems and
post-traumatic stress disorder.1 ± 5 And, it can be the
result of such physical, psychological and environmental factors and events as illness, injury or disability,
day to day worries and ®nancial concerns, and noise,
crowding or inadequate living situations.2,6
*Correspondence:
Colorado, USA
KA
Gerhart,
Craig
Hospital,
Englewood,
It is the dual role of causality and response that
makes the study of stress so complicated. The situation
is further complicated by the fact that stressors or
stressful situations alone do not necessarily lead to
adverse outcomes; not all stress is negative. Moreover,
not all stress is created by negative events. Marriage, a
new job, or the birth of a new baby, for example, are
events that, to many people, are simultaneously
positive and highly stressful.
In any case, it is reported that two characteristics
must be present for stress to have an impact on the
individual: a given situation must be felt by the
individual to be demanding or threatening, and the
individual must believe that he or she does not have
Stress in long term SCI
KA Gerhart et al
184
the resources to cope with the situation.7 ± 9 As a
result, individuals will vary widely in what they view
as stressful and how they respond to stress. It is
precisely this variation that supports arguments for
measuring perceived stress,9 ± 11 the outcome measure
selected for this research with long-term spinal cord
injury survivors.
Although disability has been reported to be a
stressor,2,6,12 the precise relationship between stress
and disability remains unclear. Fuhrer et al and
Rintala et al13,14 studied 140 spinal cord injury
survivors (100 men, 40 women), with a mean age of
37 years (standard deviation=11.5) and a mean
duration of injury of almost 11 years (SD=7.8
years). All were living in a 13-county area in and
around Houston and Galveston, Texas, USA.
Members of this sample were found to have more
stress than the general population, but the investigators were unable to attribute this increased stress to
the severity of their participants' impairments or the
extent of their disabilities. However, the extent to
which this is true in more population-based samples of
disabled people and the extent to which it applies to
longer injured SCI survivors has not been studied.
Also virtually unexamined are the factors that lead to
stress and the outcomes that result, over time, from
increased perceived stress among SCI survivors. The
purpose of this population-based research, therefore,
was to address the following questions:
(1)
(2)
(3)
What characterizes those long-injured spinal cord
injured persons who report more perceived stress
than others; what role, if any, does the extent or
severity of the disability play?
What characteristics at one point in time are
associated with future stress among SCI survivors
with disabilities of more than 20 years duration?
What are later outcomes that are associated with
earlier reports of stress among long term spinal
cord injury survivors?
Methods
Research participants were initially drawn from a
population-based sample of 412 long-term SCI
survivors meeting the following broad criteria:
(1)
(2)
(3)
(4)
residence, at the time of injury, in one of 13
counties served by the National Spinal Injuries
Centre at Stoke Mandeville Hospital in Aylesbury and by the Northwest Regional Spinal
Injuries Centre at the District General Hospital
in Southport, England;
onset of traumatic spinal cord injury prior to
December 31, 1970;
age between 15 and 55 years at the time of onset,
to include only persons with non-elderly, yet
adult-onset spinal cord injuries;
admission to one of the two participating SCI
centers within 1 year of injury;
(5)
documentation ± via the British Oce of Population Censuses and Surveys and/or the National
Health Services ± that the individual was still alive
at the study's inception in 1990.
In 1990, all of the 412 known survivors were invited
to participate in a longitudinal research e€ort studying
outcomes of long-term spinal cord injury. Two
hundred and seventy-nine agreed and underwent
comprehensive assessments at the SCI centre which
included a full battery of health, medical, and
functional evaluations conducted by a physician and
physical therapist, an extensive interview querying
them about their perceived health, quality of life,
and functional concerns, and a battery of psychometric tests in order to assess depression, perceived
stress, well-being, and life satisfaction. Two hundred
and twenty-seven participants returned for the second
round of longitudinal data collection in 1993, and in
1996 198 returned. However, it was the 187 individuals
who participated in all three of the data collection
e€orts and who had valid scores on Cohen,
Kamarack, and Mermelstein's Perceived Stress Scale9
in both 1993 and 1996 who comprised the sample
Table 1 Participant descriptions as of 1993 data collection
e€ort (n=187)
Gender
Men
Women
Age
Mean
Range
Age groups
<50 years
50 ± 59 years
60 ± 69 years
>69 years
Duration of disability
Mean
Range
Duration of disability groups
<30 years
30 ± 39 years
>39 years
Neuro grouping
Para (ASIA A, B, C)
Tetra (ASIA A, B, C)
All ASIA D*
Marital status
Married
Single
Divorced/Separated
Widowed
Employment
Working
Not working (Unemployed/Retired)
Number
Percent
161
26
86.1
13.9
53.6
37 ± 77
69
75
32
11
36.9
40.1
17.1
5.8
29.7
23 ± 49
111
60
16
59.4
32.1
8.6
93
57
37
49.7
30.5
19.8
121
45
16
5
64.7
24.1
8.6
2.7
74
113
39.6
60.4
*There were no study participants who had been classi®ed
ASIA E
Stress in long term SCI
KA Gerhart et al
185
targeted by this particular research. Table 1 describes
the participants. It is noteworthy that those who
participated did di€er somewhat from those who were
eligible but did not participate in this research. While
limited data on non-participants obviously prevented
extensive comparisons, the two groups did di€er
signi®cantly with respect to age and level of
neurologic impairment (P50.01). Participants, as a
group, were 4.8 years younger than non-participants,
and more of the participants were described as
tetraplegics at the time of their injury (24.6% versus
10.2%), while fewer had neurologically incomplete
injuries (26.7% versus 42.7%). There were no
di€erences between the two groups with respect to
gender or duration of injury.
Instruments and analyses
The Perceived Stress Scale (PSS) assesses the degree to
which individuals living in the community appraise
situations in their lives as stressful, unpredictable,
uncontrollable, and overloading. Using ®ve response
options ± `never', `almost never', `sometimes', `fairly
often', and `very often' ± the PSS asks subjects how
often they have had particular thoughts or feelings,
speci®cally during the past month. Questions are
general, rather than descriptive of speci®c situations,
and target such general feelings as nervousness, being
overwhelmed, inability to cope with life hassles, and
anger over lack of control. For example, one question
asks, `In the last month, how often have you felt that
you could not cope with all the things that you had to
do?'
Scores for the PSS can range from 0 to 56, with
higher scores indicating more perceived stress. The
PSS is reported to be reliable and to correlate with
life-event scores, depressive and physical symptomatology, utilization of health services and, social
anxiety.9 It has also been reported to be a good
predictor of health and health-related outcomes, and
has been shown to have good internal and test-retest
reliability.9 While this research e€ort utilized the 14
question PSS, it is important to note that the Texas
study described earlier utilized a ten item version of
the Perceived Stress Scale.13,14 There is also a four
item version in use.12. In addition to this study of
spinal cord injury survivors, another report of PSS
usage among persons with physical impairments
involved subjects with tempromandibular pain and
physical therapy patients being treated for various
types of joint and muscle pain.15
The long-term spinal cord injury survivors' scores
on the Perceived Stress Scale ± administered in 1993
and 1996 ± served as the focus around which all
analyses were conducted. Treated as a continuous
variable, PSS scores were associated, using Pearson
correlations, with other data, other test scores, and
outcomes to identify signi®cant relationships. One
hundred and twenty comparisons were performed
using the P50.01 signi®cance level. Assuming that
all tests were independent of each other, the chances of
four or more being signi®cant due to chance alone is
conservatively estimated at 2.5%.
The ®rst level of analyses targeted relationships
within one time period. 1993 PSS scores were
compared with other 1993 outcomes to characterize
SCI survivors with high perceived stress; the same was
done with 1996 scores and 1996 outcomes. Then,
building on the longitudinal design of this research,
analyses were conducted across time periods in an
e€ort to identify factors and phenomena that were
associated with later stress as well as those that were
associated with previous stress. First, the variables
collected during 1990 were associated with the 1993
PSS scores to identify correlates of future stress; the
same was done with 1993 variables and 1996 PSS
scores. Then, the 1993 PSS scores were associated with
1996 outcomes in order to identify the results and
longer term implications of earlier perceived stress.
Variables examined in all of these time periods
included such standard variables as gender, age, and
duration of injury, as well as the potential role of the
level and severity of injury. The latter was assessed in
®ve ways: by hours of attendant care utilized each day;
by Functional Independence Measure scores (FIM16);
by physical independence subscale scores measured by
CHART,17 by American Spinal Injury Association
(ASIA)/International Medical Society of Paraplegia
(IMSOP) motor scores;18 and, ®nally, by comparison
of three functional impairment subgroups, based on
the ASIA/IMSOP classi®cation:18
. those with functionally complete paraplegia who
typically had intact upper extremities but used
wheelchairs for their mobility (Para: ASIA ABC)
. those with functionally complete tetraplegia who
similarly used wheelchairs but also had impaired
function in their upper extremities (Tetra: ASIA
ABC)
. those with functionally incomplete SCIs at any
neurological level who had neurological sparing
such that ambulation was typically possible (All
ASIA, D)
Additionally, a range of reported medical complications and health interview responses were analyzed.
These included the existence of pressure sores, urinary
and bowel-related complaints and diagnoses, neurologic changes, joint pain, fatigue, medication concerns,
dietary issues, sexuality issues, functional decline, and
a host of others.
Also analyzed were outcomes related to adjustment
and function, as measured by the following scaled tests
and instruments:
. The Center for Epidemiological Studies Depression
Scale (CES-D19) is a 20 item scale that asks users to
report how many days during the past week they
have experienced symptoms like feeling happy or
fearful, or being unable to `shake o€ the blues'.
Stress in long term SCI
KA Gerhart et al
186
Item scores are then summed, yielding a total score
which can range from 0 ± 60, with higher scores
indicating greater depressive symptomatology. The
CES-D is widely used in community populations,
including the elderly, and it correlates signi®cantly
with clinical depression ratings, with other measures
of depression, and with actual diagnoses of
depression.20 ± 24 Developed as a tool for measuring
sample-wide characteristics of depression, its reliability has been measured using corrected split half
correlations (0.85) and an alpha coecient (0.92).25
Himmelfarb and Murrell have shown the CES-D to
have high reliability among elderly populations.26
Validity of the CES-D is reinforced by the
persistence and consistency of relationships between depression and health, gender, and marital
status across ages.27 ± 29 Within the realm of spinal
cord injury, the CES-D has been used in studies of
adjustment,30 pain,31 substance abuse,32 as well as
to assess the overall levels of depression.33,34
. The Life Satisfaction Index (LSI-Z35) assesses
morale or satisfaction with life by asking respondents to agree or disagree with 13 statements like,
`These are the best years of my life', and `The things
I do are as interesting to me as they ever were'. This
instrument was developed after item analysis of the
longer LSI-A version and correlates adequately with
the Life Satisfaction Rating.35,36 Its validity and
reliability have been studied, and have been found
to be satisfactory, particularly among elderly
populations where it has been used extensively
(Kuder-Richardson Coecient Alpha=0.079).37 ± 43
Additionally, the LSI-Z has also been used in
studies of aging and spinal cord injury.44,45
. The Index of Psychological Well-being (IPWB46)
consists of eight items, with response options of
`never', `sometimes', and `often'. Examples of
questions are: `How often do you feel bored?' and
`How often do you feel particularly excited or
interested in something?' The instrument is scored
on a seven-point scale, and scores higher than `4'
typically indicate poorer ± and less than average ±
well-being. The Index of Psychological Well Being
has been validated through its similarity in relationships to a stress-factor and its association with an
index of neurotic traits.46 While not extensively
used in spinal cord injury, a few studies have shown
that people with SCI scored similarly to Bergman's
general population sample.30,45
. The Functional Independence Measure (FIM;16),
quanti®es the extent of disability by assessing an
individual's level of physical and cognitive independence. Performed of 18 tasks in six life care areas ±
including self-care, locomotion, mobility, sphincter
control, communication, and social cognition ± is
assessed with a seven point scale describing varying
degrees in the amount of assistance or supervision
that is needed from others as the result of an
impairment. Subscores for each of the six life areas,
as well as a total score are obtained. The FIM's
ability to detect changes in independence over time
is reported to be high, and inter-rater reliability is
good among rehabilitation clinicians who administer the instrument. In a multi-center study involving
over 1000 patients, pairs of clinicians obtained a
inter-rater correlation coecient of 0.99 for the
total FIM score; for none of the six subscores was
the coecient lower than 0.97.47,48 Its face validity
and construct validity have been demonstrated in
studies with clinicians, and the FIM has been
shown to accurately predict burden of care, defined
as minutes of help received per day.47
. The Craig Handicap Assessment and Reporting
Technique (CHART17) is a tool for measuring
handicap or level of social integration for people
with disabilities. It uses six subscales which closely
parallel the domains of handicap described by the
World Health Organization49 ± orientation, physical
independence, mobility, occupation, social integration, and economic self-suciency. The instrument
contains 32 questions which attempt to quantify the
extent to which individuals ful®ll various social
roles. It is normed on a nondisabled sample such
that majority of those completing CHART attained
the maximum score of 100 in each subscale ±
indicating that they had no handicap(s). For the
total CHART score, the test-retest reliability
coecient was found to be 0.93; the reliability
coecient between subjects and their proxies was
0.83 and independent raters have established
CHART's validity. Rasch analysis further veri®ed
CHART's scaling and scoring procedures.17
Results
Mean PSS scores
First, Table 2 depicts the mean PSS scores for various
subgroups of the sample. No signi®cant (P50.01)
di€erences ± during either time period ± could be
demonstrated with respect to age, duration of injury,
gender, or severity of the impairment. However, mean
PSS scores did decrease signi®cantly between the 1993
and 1996 administrations (P50.001).
Analysis within the same time period: characteristics of
SCI survivors with higher perceived stress
As Table 3 shows, not only were there no signi®cant
correlations between stress and age or duration of
injury, but neither were there associations that could be
found linking stress levels to the extent or severity of
the SCI survivors' disability. There were no signi®cant
associations between PSS scores and ASIA Motor
scores, FIM scores, the level of physical independence
(as measured by the CHART subscale), or the number
of hours of attendant care used per day. Even when
those with tetraplegia were compared to those with
paraplegia and with functionally incomplete spinal
cord injuries, there were no di€erences.
Stress in long term SCI
KA Gerhart et al
187
With respect to physical symptoms, those with more
stress also had signi®cantly more fatigue in 1993, but
not in 1996. They also had more stomach pain and
nausea in one time period but not the other. No other
medical health complaints ± including pressure sores
and upper extremity pain ± were signi®cantly associated with PSS scores at either time period.
However, as Table 3 also shows, stress in both 1993
and 1996 was correlated with several concurrent
psychological outcomes. These included lower life
satisfaction (as measured by the LSI-Z), lower selfrated quality of life, lower psychological well-being (as
measured by the IPWB), and depressive symptoms. In
fact, in addition to the total CES-D score, each
depression subscale itself was signi®cantly related to
perceived stress, with higher levels of the latter being
associated with more a€ective symptoms, more
somatic symptoms, less well-being, and lower interpersonal a€ect.
Additionally, those who had higher occupation
subscores on CHART ± indicating that they were
more active vocationally and avocationally ± also
had higher stress scores in 1993. Moreover, the
entire CHART score itself ± a measure of handicap
and community integration ± was related to stress.
Those survivors with the highest CHART scores ±
those who were the least handicapped and seemingly
the most involved in typically expected social roles ±
reported the most stress in 1993. However, neither of
these relationships with CHART scores retained their
signi®cance in 1996. Regression analysis showed that
47% of the variance in the 1993 PSS scores could be
accounted for by the signi®cant associated variables
listed in Table 3, while in 1996, 52% of PSS score
variance was accounted for by its signi®cantly
associated variables.
Table 2 Mean PSS scores
Overall
Age groups
<50
50 ± 59
60 ± 69
469
Duration of injury
<30 years
30 ± 39 years
439 years
Gender
Men
Women
Impairment
Tetra ASIA ABC
Para ASIA ABC
All ASIA D
PSS 1993
Mean Standard
score deviation
PSS 1996
Mean Standard
score deviation
18.02
8.43
17.61
8.53
18.68
18.76
16.47
13.27
8.17
8.99
7.72
6.66
18.33
17.64
16.06
17.36
8.63
8.70
7.97
8.92
18.30
19.02
12.31
8.54
8.13
6.84
17.63
18.27
15.00
8.68
8.53
7.44
17.76
19.62
8.57
7.45
17.44
18.65
8.68
8.53
18.16
18.01
17.81
9.23
7.90
8.65
17.82
18.09
16.08
9.08
8.38
8.07
Analysis across time periods: characteristics and factors
associated with stress 3 years later
With the exception of fatigue in 1996, medical issues
and complaints ± including pressure sores and upper
extremity pain ± had no consistent relationship with
stress reported 3 years later. However, scores on the
psychological instruments were again signi®cant. Lower life satisfaction in 1990 was associated with higher
Table 3 Associations with Perceived Stress Scale score:
during the same time period
PSS 1993
correlation
Variable
Age
Duration of injury
Employed
Occupation
(CHART subscore)
Gender
Extent of disability
FIM score
Extent of disability
ASIA motor scores
Extent of disability
Para ABC
Tetra ABC
All DE
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
1993
1996
Extent of disability
hours of care
Extent of disability
(CHART physical
independence subscale
score)
1993
Fatigue
1996
1993
Upper extremity pain
1996
1993
Increased c/o stomach
1996
pain
1993
Increase c/o nausea
1996
1993
Absence/presence of
1996
pressure sore (any
grade, anywhere)
1993
Total CHART score
1996
1993
Life satisfaction
1996
(LSI-Z score)
1993
Depression
1996
(CES-D score)
Psychological Well-being 1993
1996
score
Perceived qualify of life 1993
1996
**P<0.01; ***P<0.001
±0.172
±0.161
0.148
0.191**
0.076
0.010
±0.006
70.001
0.011
70.012
0.017
0.015
0.258***
70.006
0.276***
0.110
0.081
0.216**
70.275***
0.645***
0.461***
70.353***
PSS 1996
correlation
±0.053
±0.036
0.114
0.037
0.049
±0.051
±0.032
0.056
0.017
70.089
0.037
70.049
0.185
0.040
0.109
0.180
0.123
0.048
70.374***
0.713***
0.536***
70.444***
Stress in long term SCI
KA Gerhart et al
188
perceived stress in 1993; similarly, lower life satisfaction in 1993 was related to stress in 1996. Depressive
symptoms, poorer perceived well-being, and poorer
reported perceived quality of life were also associated
with stress 3 years later, as Table 4 illustrates.
Regression analysis showed that 15% of the variance
in the 1993 PSS scores could be accounted for by the
signi®cant associated variables listed in Table 4. In
1996, 25.6% of the variance in PSS scores was
accounted for by its signi®cantly associated variables.
Analysis across time periods: the relationship of earlier
stress to outcomes 3 years later
As in the preceding analyses, psychological outcomes
as ± measured by the various instruments ± proved to be
the most signi®cant outcomes of earlier stress. First,
individuals with higher PSS scores in 1993 were likely
to have high PSS scores again 3 years later. They were
also more likely to report lower life satisfaction 3 years
later based on LSI-Z scores, lower well-being as
measured by the Index of Psychological Well-Being,
and more depressive symptomatology as re¯ected by
higher total CES-D scores. People with more stress in
1993 also used more alcohol in 1996. Interestingly,
stress in 1993 was also associated with higher economic
Table 4 Previous characteristics signi®cantly associated with
stress 3 years later
PSS 1993
correlation
Variable
Life satisfaction
(LSI-Z)
Psychological wellbeing
Depressive (total
CES-D score)
Perceived quality
of life
Fatigue
Stomach pain
Nausea
1990
1993
1990
1993
1990
1993
1990
1993
1990
1993
1990
1993
1990
1993
70.207**
0.389***
Not administered
70.167
PSS 1996
correlation
70.216**
0.350***
0.468***
70.238**
0.159
0.270***
0.190
0.234**
0.149
0.220**
**P50.01 ***P50.001
Table 5
The relationship of stress in 1993 to 1996 outcomes
Outcome in 1996
PSS 1993
correlation
0.599***
Perceived stress scale
70.156
Life satisfaction (LSI-Z)
0.362***
Psychological well-being
0.369***
Depressive symptoms (total CES-D score)
0.238**
Economic self-suciency (CHART subscale)
**P50.01 ***P50.001
self-suciency in 1996, as measured by the CHART
subscale. Total CHART scores also were higher,
indicating more stress among those who became more
highly integrated into and active within their communities. Finally, of all the health issues and physical
complications ± including all those reported previously ± only fatigue in 1996 was signi®cantly associated with previous stress.
Discussion
Perhaps the most important contribution of this
research ± and perhaps its most signi®cant implications
for clinicians ± is three relationships it appears to
demonstrate:
(1)
(2)
(3)
the non-relationship between stress and the
severity of the disability among long-term spinal
cord injury survivors.
the seemingly strong relationship between stress
and other psychological outcomes, particularly
depression, life satisfaction, and quality of life.
the lack of strong relationships between stress
and medical outcomes, despite the plethora of
scienti®c research and reports in the popular
press on the impact of stress on physical health.
All three of these ®ndings are supported by recently
published research with an American sample of SCI
survivors. In two reports from the Baylor College of
Medicine Life Status Study, a similarly non-signi®cant
relationship between stress and severity of the
impairment or disability was found. Involving 661
Texans with a mean age of 37 and a mean duration of
injury of almost 11 years, this study found no PSS
score di€erences with respect to participants' level and
completeness of injury, their ASIA Total Motor Index
scores, self-reported FIM scores, or the amount of
personal assistance they received.13,14 Nonetheless,
relationships were identi®ed between stress and
measures of life satisfaction, depressive symptomatology, and self-assessed health, while no relationships
were found between stress and two key SCI-speci®c
medical complications: urinary tract infections and
pressure sores.14 The absence of a relationship between
stress and pressure sores, perhaps the most prevalent
and certainly one of the most devastating complications faced by long-term spinal cord injury survivors is
particularly conspicuous in both studies and seems
almost counter-intuitive. Further research seems
indicated.
Two other medical correlates of stress identi®ed in
the British population-based sample also seem to merit
further study. Both fatigue and bowel and stomachrelated complaints exhibited some inconsistent association with stress. Though these may not be as dramatic
as the hypertension or cardiovascular disease associations found in the nondisabled population, both
fatigue and gastrointestinal problems are concerns
that are prevalent in long-term SCI survivors and as
Stress in long term SCI
KA Gerhart et al
189
such may merit further monitoring for their impact on
survivors' perceptions of themselves and their ability
to function e€ectively. Further study is also needed to
better understand why some outcomes were significantly associated with stress across one time period,
but not the other.
There are several limitations to this study. First, it is
important to keep in mind that this research identi®ed
only relationships and associations. It's dicult to
conclude that stress in these British SCI survivors is
the cause of particular concomitant or later outcomes,
or that particular symptoms or characteristics were
themselves the direct causes of later reported stress.
Any of the identi®ed stress correlates could be the
result of a more complex interrelationship among
numerous variables, or of some other, unstudied
variable. Indeed, despite the signi®cant correlations
identi®ed, there is still a large amount of variance
unaccounted for by the variables assessed in this
study. Perhaps unmeasured variables unrelated to SCI
play a role in the perceived stress measured by this
research.
Additionally, it is not known to what extent the
study participants' perceived stress was confounded by
psychological distress. This has been a limitation of
the PSS itself, according to its authors.12 It is not
known how directly the SCI survivors' stress, as well
as their life satisfaction, well being, and depressive
symptomatology, were in¯uenced by or were a
function of their own self-perceptions, adjustment
and personal coping strategies. While others have
studied the role of coping strategy on distress8,50 this
was not examined in our population.
In addition, the extent to which these ®ndings are
generalizable is not known. With the exception of
the recently published ®ndings of the Baylor College
of Medicine Life Status Study,13,14 few, if any, other
uses of the PSS with SCI survivors are reported.
That the British and Baylor studies were similar with
respect to the three main ®ndings described above is
encouraging. However, other questions remain, and
it must be kept in mind that while the British sample
as a whole was population-based, the group of longterm SCI survivors who chose to participate in this
research contained a disproportionate number of
individuals who are younger and who have more
severe injuries.
Perhaps the most important question relates to how
mean stress levels compare across groups of disabled
and nondisabled persons. Direct comparisons with the
Baylor group are limited by the fact that di€erent
versions of the PSS were used. However, if scores are
prorated to adjust for the use of 14 and 10 item
instruments, it appears that the American sample ±
which is considerably younger ± had higher mean stress
scores than the British group. Though such a score
proration is clearly an oversimpli®cation, it does
further substantiate one ®nding that has appeared
repeatedly: stress levels decrease with age. In the
British sample, younger participants had higher stress
scores; in both samples, mean stress scores dropped
when the PSS was administered 3 years later; and, in
the older British sample mean scores were lower than
in the younger American group. Even in Cohen and
Williamson's general population sample, scores decreased as age increased.12
What about more direct comparisons with the
general population? Do SCI survivors have more
stress than nondisabled people? Though the Baylor
study concluded that its SCI survivors did have more
perceived stress than their general population counterparts,13,14 this was less clear in the older British
sample. Compared with Cohen and Williamson's
report,12 stress scores for the British SCI participants
actually seemed to be very slightly lower than for the
general population. However, the extent to which this
was a€ected by di€erences in how age groups were
de®ned, by the use of di€erent versions of the PSS, by
cross-cultural di€erences ± or whether it truly re¯ects
an increasing similarity in stress levels as both disabled
and nondisabled people age ± cannot be determined.
Clearly, only further research which samples SCI
survivors of many ages and durations of injury and
includes matched contemporary nondisabled controls,
will de®nitively answer this question.
References
1 Kasl SV. Stress and Health. Ann Rev Public Health 1984; 5: 319 ±
341.
2 Butt LM, Lanig IS. Stress Management. In, Lanig IS, Chase TM,
Butt LM, Hulse KL, Johnson KMM (eds). A Practical Guide to
Health Promotion after Spinal Cord Injury. Gaithersburg, MD:
Aspen Publishers, 1996.
3 Dohrenwend BP, Dohrenwend BS. Social status and psychological disorder: a causal-injury. New York: John Wiley and Sons,
1969.
4 Goldberg EL, Comstock GW. Life events and subsequent illness.
American Journal of Epidemiology 1976; 194: 146 ± 156.
5 Myers JK et al. Life events and mental status: a longitudinal
study. Journal of Health and Social Behavior 1972; 13: 398 ± 406.
6 Holmes TH, Rahe RH. The social readjustment rating scale.
Journal of Psychosomatic Research 1967; 11: 213 ± 218.
7 Buelow JM. A correlational study of disabilities, stressors, and
coping methods in victims of multiple sclerosis. Journal of
Neuroscience Nursing 1991; 23: 247 ± 251.
8 Folkman S, Lazarus RS. An analysis of coping in a middle-aged
community sample. Journal of Health and Social Behaviour 1980;
21: 219 ± 239.
9 Cohen S, Kamarack T, Mermelstein RA. Global measure of
perceived stress. Journal Health and Social Behaviour 1983; 24:
385 ± 396.
10 Lazarus RS. Psychological Stress and the Coping Process. New
York: McGraw-Hill, 1966.
11 Vingerhoets AJJM, Marcelissen FHG. Stress research: its present
status and issues for future developments. Soc Sci Med 1988; 26:
279 ± 291.
12 Cohen S, Williamson GM. Perceived stress in a probability
sample of the United States. In Oskamp S, Spacapan S (ed). The
Social Psychology of Health: Claremont Symposium Applied
Social Psychology. Newbury Park, CA: Sage Publications, 1988.
13 Fuhrer M, Rintala D, Hark K. Interrelationships among
psychological stress and the components of disablement for
people with spinal cord injury. IRMA VIII: The 8th World
Congress of the International Rehabilitation Medicine Association. Bologna, Italy: Monduzzi Editore SpA, 1997.
Stress in long term SCI
KA Gerhart et al
190
14 Rintala DH, Hart KA, Fuhrer MJ. Perceived stress in individuals
with spinal cord injury. In Kratoski DM, Nosek MA, Turk MA
(eds). Women with Physical Disabilities: Achieving and
Maintaining Health and Well-being. Baltimore, MD: Paul H.
Brookes Publishing, 1997.
15 Schnurr RF, Brooke RI, Rollman GB. Psychosocial correlates of
tempromandibular joint pain and dysfunction. Pain 1990; 42:
153 ± 165.
16 Guide for the Uniform Data Set for Medical Rehabilitation
(including the FIM Instrument) 1996; Bu€alo, NY: State
University of New York at Bu€alo.
17 Whiteneck GG et al. Quantifying handicap: a new measure of
long-term rehabilitation outcomes. Archives of Physical Medicine
& Rehabilitation 1992; 73: 519 ± 526.
18 Ditunno JF, Young W, Donovan WH, Creasey G. The
international standards booklet for neurological and functional
classi®cation of spinal cord injury. Paraplegia 1994; 32: 70 ± 80.
19 Radlo€ LS. The CES-D Scale: a self-report depression scale for
research in the general population. Applied Psychological
Measures 1977; 3: 385 ± 401.
20 Radlo€ LS, Teri L. Use of the Center for Epidemiological Studies
Depression Scale with older adults. Clinical Gerontology 1986; 5:
119 ± 136.
21 Hertzog C et al. Measurement properties of the Center for
Epidemiological Studies Depression Scale (CES-D) in older
populations. Journal of Consulting Clinical Psychology 1986; 2:
64 ± 72.
22 Roberts RE, Vernon SW. The Center for Epidemiological
Studies Depression Scale: Its use in a community sample.
American Journal of Psychiatry 1983; 140: 41 ± 46.
23 Weismann MM et al. Symptom patterns in primary and
secondary depression L A comparison of primary depressives
with depressed opiate addicts, alcoholics, and schizophrenics.
Arch Gen Psychiatry 1997; 34: 854 ± 862.
24 Husaini BA et al. Depression in rural communities: Validating
the CES-D scale. Journal of Community Psychology 1980; 8: 137 ±
146.
25 Radlo€ LS. The CES-D Scale: A self-report depression scale for
research in the general population. App Psychol Meas 1977; 3:
385 ± 401.
26 Himmelfarb S, Murrell SA. Reliability and validity of ®ve mental
health scales in older persons. Journal of Gerontology 1983; 38:
333 ± 339.
27 Radlo€ LS. Risk factors for depression: What to we learn from
them? from Mental health of women: facto and ®ction. Belle D
and Salasin S (eds). New York: Academic Press, 1980.
28 Sayetta RB. Basic data on depressive symptomatology: United
States, 1974 ± 1975. Washington DC. DHEW Publication #80 ±
1666.
29 Goldberg E, Van Natta P, Comstock G. Depressive symptom,
social networks and social supports of elderly women. American
Journal of Epidemiology 1985; 121: 448 ± 456.
30 Schulz R, Decker S. Long-term adjustment to physical disability:
the role of social support, perceived control, and self-blame.
Journal of Personality and Social Psychology 1985; 1162 ± 1172.
31 Cairns DM, Adkins RH, Scott MD. Pain and depression in acute
traumatic spinal cord injury: origins of chronic problematic pain?
Arch Phys Med Rehabil 1996; 77: 329 ± 335.
32 Young ME et al. Alcohol and marijuana use in a communitybased sample of persons with spinal cord injury. Arch Phys Med
Rehabil 1995; 76: 525 ± 532.
33 Fuhrer MJ et al. Depressive symptomatology in persons with
spinal cord injury who reside in the community. Arch Phys Med
Rehabil 1993; 74: 255 ± 260.
34 Daverat P et al. The long term outcome in 149 patients with
spinal cord injury. Paraplegia 1995; 33: 665 ± 668.
35 Wood V, Wylie ML, Sheafor B. An analysis of a short self-report
measure of life satisfaction: correlation with rater judgments.
Journal of Gerontology 1969; 24: 465 ± 469.
36 Neugarten BL, Havighurst RJ, Tobin SS. The measurement of
life satisfaction. Journal of Gerontology 1961; 16: 134 ± 143.
37 Baiyewu O, Jegede RO. Life satisfaction in elderly Nigerians:
reliability and factor composition of the life satisfaction Index Z.
Age & Ageing 1992; 21: 256 ± 261.
38 Stock WA, Okun MA. The construct validity of life satisfaction
among the elderly. Journal of Gerontology 1982; 37: 625 ± 627.
39 Bullard-Poe L, Powell C, Mulligan T. The importance of
intimacy to men living in a nursing home. Arch Sex Behav
1994; 23: 231 ± 236.
40 Clark F et al. Occupational therapy for independent-living older
adults. A randomized controlled trial. JAMA 1997; 278: 1321 ±
1326.
41 Rubin CD, Sizemore MT, Loftis PA, de Mola NL. A
randomized, controlled trial of outpatient geriatric evaluation
and management in a large public hospital. J Am Geriatr Soc
1993; 41: 1023 ± 1028.
42 Kozma A, Stones MJ. The measurement of happiness:
development of the Memorial University of Newfoundland
Scale of Happiness (MUNSH). Journal of Gerontology 1980;
35: 906 ± 912.
43 Chat®eld WF. Economic and sociological factors in¯uencing life
satisfaction of the aged. Journal of Gerontology 1977; 32: 593 ±
599.
44 Weitzenkamp DA et al. Spouses of spinal cord injury survivors:
the added impact of caregiving. Arch Phys Med Rehabil 1997; 78:
822 ± 827.
45 Whiteneck GG et al. Mortality, morbidity, psychosocial outcomes of persons spinal cord injured more than 20 years ago.
Paraplegia 1992; 30: 617 ± 630.
46 Berkman PL. Measurement of mental health in a general
population survey. American Journal of Epidemiology 1971; 94:
105 ± 111.
47 Granger CV, Fiedler RC. The measurement of disability. In
Fuhrer MJ (ed). Assessing Medical Rehabilitation Practices: the
Promise of Outcomes Research. Baltimore, MD: Paul H. Brookes
Publishing, 1997.
48 Dahmer ER et al. A model systems database for traumatic brain
injury. Journal of Head Trauma Rehabilitation 1993; 8: 12 ± 25.
49 World Health Organization. International classi®cation of
impairments, disabilities, and handicaps: a manual of classification relating to the consequences of diseases. Geneva: World
Health Organization. 1980.
50 Kennedy P, Lowe R, Grey N, Short E. Traumatic spinal cord
injury and psychological impact: a cross-sectional analysis of
coping strategies. British Journal of Clinical Psychology 1995; 34:
627 ± 639.